Causes and consequences of contrasting genetic structure in sympatrically growing …hirc.botanic.hr/EpiSalvia/Docs/Radosavljevic_2015_AOB... · 2015. 11. 15. · tions have been

Research Article

Causes and consequences of contrasting genetic structurein sympatrically growing and closely related speciesIvan Radosavljevic1, Zlatko Satovic2 and Zlatko Liber1*1 Faculty of Science, Department of Botany, Division of Biology, University of Zagreb, Marulicev trg 9a, HR 10000 Zagreb, Croatia2 Faculty of Agriculture, Department of Seed Science and Technology, University of Zagreb, Svetosimunska cesta 25, HR 10000 Zagreb,Croatia

Received: 2 June 2015; Accepted: 14 August 2015; Published: 2 September 2015

Associate Editor: F. Xavier Pico

Citation: Radosavljevic I, Satovic Z, Liber Z. 2015. Causes and consequences of contrasting genetic structure in sympatrically growingand closely related species. AoB PLANTS 7: plv106; doi:10.1093/aobpla/plv106

Abstract. Gene flow, natural selection and genetic drift are processes that play a major role in shaping the geneticstructure of natural populations. In addition, genetic structures of individual populations are strongly correlated withtheir geographical position within the species distribution area. The highest levels of genetic variation are usuallyfound in the centre of a species’ distribution and tend to decrease beyond that point. Additionally, narrowly endemictaxa are expected to be characterized by lower levels of genetic variation than their widespread congeners. To under-stand the historical circumstances that shape populations of sympatric and closely related taxa, microsatellite mar-kers were used, while populations of the three closely related and sympatric Mediterranean Salvia species (S. officinalisL., S. fruticosa Mill. and S. brachyodon Vandas) served as a study model. In the populations of widespread S. officinalis,located in the central parts of this species’ distribution area, no population genetic disturbances were detected. Thenarrow endemic S. brachyodon showed heterozygote excess, clonal reproduction and a genetic bottleneck. Becausethe genetic bottleneck was likely caused by the disappearance of suitable open-type habitats, the recent wildfire thatcleared the terrain probably saved the S. brachyodon population from gradual deterioration and extinction. At thesame time, clonal reproduction could serve as a valuable mechanism in the preservation of genetic variability. Theresults of the disjunct S. fruticosa population indicated heterozygote deficiency, inbreeding, hybridization withS. officinalis and population expansion. The hybridization with S. officinalis along with the abandonment of theagro-pastoral system are likely the main drivers of the strong expansion of S. fruticosa in the studied location. Asmany relevant findings and conclusions regarding historical and contemporary demography of individual populationsor species can be reached only through their comparison with closely related taxa, this study demonstrates the import-ance and advantages of such a multi-species approach.

Keywords: Hybridization; Mediterranean; population bottleneck; population genetics; Salvia; SSR.

IntroductionThere are three processes that are believed to play majorroles in shaping the genetic structure of natural popula-tions: gene flow, natural selection and genetic drift

(Lesica and Allendorf 1995; Eckert et al. 2008). Numerouspast events, either natural or human-mediated, can havesubstantial effects on the current intra- and interspecificgenetic variability. Interspecific hybridization, genetic

* Corresponding author’s e-mail address: [email protected]

Published by Oxford University Press on behalf of the Annals of Botany Company.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.

AoB PLANTS www.aobplants.oxfordjournals.org & The Authors 2015 1

by guest on Novem

ber 15, 2015http://aobpla.oxfordjournals.org/

Dow

nloaded from

http://creativecommons.org/licenses/by/4.0/



http://aobpla.oxfordjournals.org/

bottlenecks, the founder effect, inbreeding depressionand rapid population expansion are just some of theseeffects. Genetic drift events, such as the founder effectand bottlenecks, often lead to increased rates of inbreed-ing, a loss of heterozygosity and the fixation of deleteri-ous alleles (Luikart et al. 1998). With the increasedprobability of the homozygous expression of recessivedeleterious alleles, the risk of extinction increases formany species undergoing population bottlenecks (Luikartet al. 1998). Therefore, the early detection of a bottleneckevent is of central interest to conservation biologists.Among other mechanisms, hybridization plays an import-ant role in the evolution of many taxonomic groups andcan result in the formation of entirely new species.Hybridization can be of natural or anthropogenic origin;it occurs with or without introgression and can also leadto adaptive evolution and diversification via heterosis orthe production of novel genotypes capable of expressingsuperior phenotypic traits (Stebbins 1959; Arnold 1997;Rieseberg et al. 2003; Seehausen 2004).

Differences and contrasting patterns in genetic struc-tures across core, edge and disjunct populations are thesource of never ending debates and research (e.g. Gaston2003; Sagarin et al. 2006; Samis and Eckert 2007; Eckertet al. 2008). It is generally assumed that the highest abun-dance and the highest levels of genetic variability can befound in the geographical centre of species distribution(Surina et al. 2014). As we move away from the species dis-tribution centre, population sizes gradually decrease asdoes genetic diversity. Similarly, a disjunct population isexpected to be characterized by impoverished geneticstructure as a direct consequence of its origin associatedwith some sort of bottleneck event followed by theabsence of inter-population gene flow (Meeus et al.2012). The origin of disjunct populations is usually linkedto one of three possible causes: (i) human introduction,(ii) long distance dispersal and (iii) isolation after speciesrange contractions due to either climatologically oranthropogenically induced reasons (Meeus et al. 2012).

The Mediterranean basin is known not only for its gla-cial refugium areas that have played a crucial role in post-glacial recolonizations of Europe (e.g. Taberlet et al. 1998;Hewitt 1999, 2001; Michaux et al. 2003) but also for thelong-lasting human presence and the influence ofhumans on local ecosystems that can be dated back asfar as 50 000 years ago (Naveh and Dan 1973). At thebeginning of the Holocene, some 10 000 years ago, thisanthropogenic pressure on living systems became moreintense as plants and animals were domesticated, andgradually, a sustainable agro-pastoral system developed(Miller 1992; Zeder and Hesse 2000; Blondel 2006). Conse-quently, the vast majority of present-day Mediterraneanlandscapes and accompanying species communities

serve as examples of ‘coevolution’ between nature andhumans that have lasted since the last glacial period.During this time, human activities and practices haveshaped the complex interaction between human soci-eties and the local ecosystems and have subsequentlycontributed to maintaining high levels of biological diver-sity (Blondel 2006; Fonderflick et al. 2010).

With over 900 species distributed worldwide, the genusSalvia is by far the largest and most diverse genus in theLamiaceae family (Walker et al. 2004). In Europe, 36 taxaare described and grouped into seven sections (Hedge1972). Among them is the section Salvia, which comprises13 taxa at species rank (Hedge 1972), of which Salvia offi-cinalis L. (common sage) (Fig. 1) and S. fruticosa Mill. (Greeksage) (Fig. 1) are best known because of their economicallyrelevant high proportions and quality of essential oil(Putievsky et al. 1990; Dudai et al. 1999). As a typical mem-ber of indigenous flora, S. officinalis naturally grows alongthe eastern Adriatic coast (Pignatti 1982), in the centraland southern Apennines (Lucchese et al. 1995; Cutiniet al. 2007; Di Pietro 2011), and a few relict populationscan be found in continental parts of the Balkan Peninsula(Vasic 1997). The range of S. fruticosa extends from north-eastern Libya to Sicily and southern Italy through thesouthern part of the Balkan Peninsula to western Syria(Hedge 1982; Karousou et al. 2000), but this species hasbeen introduced to many Mediterranean countriesthroughout history (i.e. Malta, Spain, Portugal) by the Phoe-nicians and the Greeks (Karousou et al. 2000). In the cen-tral Adriatic region, an isolated population of Greek sagecan be found on the Island of Vis (Croatia) (Flora CroaticaDatabase; http://hirc.botanic.hr/fcd). Because this popula-tion is located �500 km from the nearest population ofthe same species in southern parts of the Republic of Alba-nia, its origin is unclear, and it is uncertain whether thispopulation is indigenous or naturalized, especially consid-ering that in the 4th century BC, a well-organized Greekcolony called Issa was founded at this location.

In addition to S. officinalis and S. fruticosa, in the coastalAdriatic region, a single additional member of the sectionSalvia can be found: Salvia brachyodon Vandas, or short-toothed sage (Fig. 1). Unlike common sage and Greeksage, S. brachyodon is a narrowly endemic species andcan be found in only two localities: St. Ilija, the highestpeak of the Peljesac Peninsula (Croatia), and Mt. Orjen,located 120 km to the southeast, in the border area of Bos-nia and Hercegovina and Montenegro (Barbalic 1956;Abadzic and Silic 1982; Silic 1984). Because of its very lim-ited distribution, habitat fragmentation, obvious ecologicalsuccession and other potential environmental threats (e.g.wildfires), short-toothed sage is believed to be a near-threatened species in Croatia (Flora Croatica Database;http://hirc.botanic.hr/fcd) and a highly vulnerable species

2 AoB PLANTS www.aobplants.oxfordjournals.org & The Authors 2015

Radosavljevic et al. — The contrasting genetic structure in sympatric and closely related species

by guest on Novem


Dow

nloaded from

http://hirc.botanic.hr/fcd







in Montenegro (Petrovic et al. 2008). Unfortunately, to thebest of our knowledge, there have been no studies focusedon this species for any reason other than its essential oilcontent (Savikin-Fodulovic et al. 2002; Tzakou et al.2003), so numerous questions regarding the ecologicalstatus and genetic condition of this species’ populationshave yet to be answered.

To perform comparative population genetic research onselected species growing in an area characterized by intri-guing climatological and ecological history and understrong and long-lasting anthropogenic influence (Blondel2006; Zeder 2008), recently developed SSR markers(Molecular Ecology Resources Primer Development Consor-tium et al. 2010; Radosavljevic et al. 2011, 2012) wereapplied. The goals of this study were to investigate (i) theconsequences of the bottleneck effect and asexual repro-duction on population genetic structure, (ii) which popula-tion genetic mechanisms could be of help in bypassinginbreeding depression, (iii) how population expansion canbe detected on a genetic level and (iv) the possible conse-quences of inter-species hybridization. We hypothesizedthat: (1) selected populations of S. officinalis from thegeographical centre of the distribution range are charac-terized by high levels of genetic diversity, are not sufferingfrom any significant population genetic disturbances (e.g.accumulating mutations, genetic drift, inbreeding, popu-lation bottleneck etc.) and can consequently serve as a‘control group’ in contrast to populations of the othertwo species, S. brachyodon and S. fruticosa), (2) a popula-tion of narrow endemic and endangered S. brachyodon ischaracterized by lower levels of genetic diversity than itswidespread congener, i.e. S. officinalis, (3) a presumablynon-native and disjunct population of S. fruticosa is gen-etically severely impoverished and suffers from popula-tion bottleneck, inbreeding and the absence of gene flow.

MethodsLeaf tissue from a total of 96 individuals from four popula-tions (24 individuals per population) was collected intwo localities: Peljesac Peninsula and the Island of Vis.S. officinalis was collected from both sites, S. brachyodonfrom the Peljesac Peninsula and S. fruticosa from the Islandof Vis. It is important to note that in the studied locations,S. officinalis occurs sympatrically with both taxa and evenhas an overlapping flowering period with S. fruticosa(late April through early May). The flowering period ofS. brachyodon occurs in August and does not overlapwith that of S. officinalis (I. Radosavljevic, Z. Satovic andZ. Liber, pers. obs.). Voucher specimens of studied popula-tions have been deposited in the Herbarium Croaticum(ZA) of the Department of Botany, Faculty of Science,University of Zagreb.

Figure 1. (A) Salvia officinalis in full bloom in the Island of Vis, (B)S. fruticosa in abandoned olive groves and vineyards outside thelocal settlement in the Island of Vis, (C) S. brachyodon in PeljesacPeninsula, in the burned out area. (D) Excavated S. brachyodonplant showing underground stolons.



by guest on Novem


Dow

nloaded from


Genomic DNA was extracted from silica gel-dried leaftissue by using a GenElute Plant Genomic DNA MiniprepKit (Sigma-Aldrich, St. Louis, MO, USA). A microsatelliteanalysis using 13 sequence-tagged SSR loci (Table 1) wasperformed according to Radosavljevic et al. (2012). Allmicrosatellite markers were isolated from S. officinalisand tested for cross-amplification in S. fruticosa andS. brachyodon, but with limited success (Radosavljevicet al. 2011, 2012). Due to this difficulty, we were unableto perform the entire analysis using a common set ofSSR markers. Instead, we used three sets with the max-imum possible number of common markers for each ana-lysed combination of species. SSR amplification reactionswere performed in a total volume of 20 mL containing10× PCR buffer, 1.5 mM MgCl2, 0.2 mM of each dNTP,0.075 mM TAIL FOR primer, 0.2 mM TAIL REV primer,0.2 mM M13 primer and 0.5 U of Taq HS polymerase (TakaraBio Inc., Shiga, Japan). The amplification was performedwith a GeneAmp PCR System 9700 (Applied Biosystems,Foster City, CA, USA) using a two-step PCR protocol withan initial touchdown cycle. The cycling conditions wereas follows: 94 8C for 5 min; five cycles of 45 s at 94 8C,30 s of annealing, beginning at 60 8C and lowered by 1 8Cin each cycle, and 90 s at 72 8C; 25 cycles of 45 s at 94 8C,30 s at 55 8C, and 90 s at 72 8C; and an 8-min extensionstep at 72 8C. The amplification products were run on anABI 3730XL analyzer (Applied Biosystems) and the resultswere analysed with GeneMapper 4.0 software (AppliedBiosystems).

GENEPOP 4.0 (Raymond and Rousset 1995) was used toestimate population genetic parameters (the averagenumber of alleles per locus, Na; the observed heterozygos-ity, HO; the expected heterozygosity or gene diversity, HE;and inbreeding coefficient, FIS) and to test the population

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Table 1. Population genetic parameter estimates for eachmicrosatellite locus surveyed in four populations of S. officinalis,S. brachyodon and S. fruticosa. Na, number of alleles; HO, observedheterozygosity; HE, expected heterozygosity; FIS, inbreedingcoefficient. Significant deviations from Hardy–Weinberg proportionsafter sequential Bonferroni corrections: ***significance at the 0.1 %nominal level; **significance at the 1 % nominal level; *significanceat the 5 % nominal level. N, number of individuals; G, number ofgenotypes; Sb, S. brachyodon population; So1, S. officinalis populationfrom the Peljesac Peninsula; So2, S. officinalis population from theIsland of Vis and Sf—S. fruticosa population.

Locus name Na HO HE FIS Sign.

Peljesac Peninsula—S. brachyodon (N ¼ 24, G ¼ 22)

SoUZ001 4 0.750 0.691 20.085 ns

SoUZ002 3 0.667 0.548 20.217 ns

SoUZ005 5 0.826 0.787 20.050 ns

SoUZ006 9 0.792 0.861 0.080 ns

SoUZ007 9 0.917 0.861 20.065 ns

SoUZ008 3 0.833 0.659 20.266 ns

SoUZ011 6 0.917 0.788 20.163 ns

SoUZ014 5 0.783 0.793 0.013 ns

Mean 5.50 0.811 0.748 20.084 *

Peljesac Peninsula—S. officinalis (N ¼ G ¼ 24)

SoUZ001 16 0.875 0.909 0.037 ns

SoUZ003 9 0.833 0.842 0.010 ns

SoUZ007 6 0.542 0.536 20.010 ns

SoUZ009 10 1.000 0.817 20.224 *

SoUZ011 19 0.833 0.955 0.127 ns

SoUZ013 7 0.792 0.756 20.047 ns

SoUZ014 15 0.958 0.913 20.050 ns

SoUZ020 7 0.478 0.756 0.367 ***

Mean 11.13 0.790 0.811 0.026 *

Island of Vis—S. officinalis (N ¼ G ¼ 24)

SoUZ001 10 0.875 0.828 20.057 ns

SoUZ003 7 0.583 0.652 0.106 ns

SoUZ007 6 0.783 0.717 20.091 ns

SoUZ009 4 0.833 0.667 20.250 ns

SoUZ011 8 0.750 0.686 20.094 ns

SoUZ013 6 0.826 0.760 20.087 ns

SoUZ014 8 0.870 0.840 20.035 ns

SoUZ020 6 0.208 0.395 0.473 **

Mean 6.88 0.714 0.692 20.033 ns

Island of Vis—S. fruticosa (N ¼ G ¼ 24)

SoUZ003 5 0.550 0.713 0.229 ns

Continued

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Table 1. Continued

Locus name Na HO HE FIS Sign.

SoUZ005 3 0.143 0.224 0.362 ns

SoUZ007 3 0.238 0.219 20.087 ns

SoUZ009 2 0.143 0.136 20.053 ns

SoUZ013 3 0.500 0.567 0.118 ns

SoUZ014 2 0.191 0.176 20.081 ns

SoUZ016 10 0.619 0.707 0.125 ns

SoUZ020 3 0.095 0.094 20.013 ns

Mean 3.88 0.307 0.351 0.125 *

P (Sb/So1) 0.010 0.635 0.293

P (Sf/So2) 0.012 0.005 0.016

P (So1/So2) 0.056 0.370 0.074



by guest on Novem


Dow

nloaded from


genotypic frequencies across all loci for their conformity toHardy–Weinberg expectations (HWE) (multilocus test).Estimates of Na, HO and HE in each population were com-pared using the Kruskal–Wallis test in SAS version 8.02(SAS Institute Inc., Cary, NC, USA). MICRO-CHECKER v.2.2.3(van Oosterhout et al. 2004) was used to check for scoringerrors caused by stutters or large-allele dropouts and toestimate null-allele frequencies. BOTTLENECK program ver-sion 1.2.02 (Cornuet and Luikart 1996; Piry et al. 1999) wasused to test for evidence of recent bottleneck events on thebasis of this theoretical expectation. The expected genediversity (HE) was compared with the expected gene diver-sity at mutation-drift equilibrium (HEQ) and calculated fromthe observed number of alleles under an intermediate two-phase model (TPM) assuming 30 % infinite allele model(IAM) and 70 % stepwise mutation model (SMM) (Di Rienzoet al. 1994). Based on the number of loci in our dataset, theWilcoxon signed-rank test (Luikart et al. 1998) was chosenfor the statistical analysis of heterozygote excess or defi-ciency, as recommended by Piry et al. (1999). The geneticdistances between pairs of S. officinalis, S. brachyodonand S. fruticosa samples were calculated based on eightmicrosatellite loci by using the proportion of shared-allelesdistances, Dpsa (Bowcock et al. 1994) as implemented inMICROSAT (Minch et al. 1996).

To depict reticulate relationships between S. officinalisand S. fruticosa populations, a NeighborNet diagram wasproduced from a pairwise distance matrix with SplitsTree4 (Huson and Bryant 2006).

To determine the power of the marker set to discrimin-ate among clones, the unbiased probability of identity(PIunbiased; Waits et al. 2001) and the PI for a populationcomposed only by sibs (PIsibs; Evett and Weir 1998; Taberletand Luikart 1999) was estimated for each locus by usingGIMLET software (Valiere 2002). Furthermore, to assessthe probability that two samples sharing a multilocusgenotype originate from distinct sexual reproductiveevents (i.e. from different zygotes, thus being differentgenets) we calculated Psex and Psex( f ) using the round-robin method of Parks and Werth (1993) as describedby Arnaud-Haond et al. (2007) and implemented inGenClone ver. 2.0 (Arnaud-Haond and Belkhir 2007).

To identify putative hybrids between S. fruticosa andS. officinalis, a model-based clustering method wasapplied on multilocus microsatellite data using the soft-ware STRUCTURE ver. 2.3.3 (Pritchard et al. 2000). Tenruns per cluster (K ) ranging from 1 to 6 were carried outon the Isabella computer cluster at the University of Zag-reb, University Computing Centre (SRCE). Each run con-sisted of a burn-in period of 200 000 steps followed by106 MCMC (Monte Carlo Markov Chain) replicates assum-ing admixture model and correlated allele frequencies.No prior information was used to define the clusters.

The most likely number of clusters (K ) was chosen bycomparing the average estimates of the likelihood ofthe data, ln[Pr(X|K )], for each value of K (Pritchard et al.2000), as well as by calculating an ad hoc statistic DK,based on the rate of change in the log probability ofdata between successive K values as described by Evannoet al. (2005). The program STRUCTURE HARVESTER v0.6.92was used to process the STRUCTURE results files (Earl andvanHoldt 2012).

The Bayesian method implemented by NEWHYBRIDS1.1. (Anderson and Thompson 2002) was used to assignindividuals into six classes: two pure (parental S. fruticosaand S. officinalis) and four hybrids (F1, F2 and backcrosseswith the parental populations). The program was run with-out any prior information about the hybrid status ofcollected individuals, and with the uniformative Jeffreysprior option for both mixing proportions and allele frequen-cies. The results were based on the average of five inde-pendent runs consisting of a burn-in phase of 100 000steps, and 500 000 MCMC sweeps. Following the sugges-tions of Anderson and Thompson (2002), individual gen-etic assignment to classes was based on a minimumposterior probability threshold (Tq) of 0.50.

ResultsThe S. officinalis population from the Island of Vis showedno significant deviations from HWE, while the populationfrom the Peljesac Peninsula was characterized by an over-all FIS value that was significantly different from zero (FIS ¼

0.026, P ¼ 0.029). However, a closer inspection revealedthat this difference was largely explained by a highly sig-nificant FIS value at only one locus (SoUZ020). No signifi-cant differences were noted when comparing the overallgenetic parameters of these populations (Table 1). Forboth populations, the Wilcoxon test yielded balancedresults under the TPM model, with no significant patternsof either excess or deficiency in the heterozygosity(Table 2). Low differentiation levels between S. officinalispopulations were observed (Fig. 2).

In S. brachyodon, the observed heterozygosity wasnoticeably higher than the expected (0.811 vs. 0.748,respectively), and the FIS value was significantly differentfrom zero (FIS ¼ 20.084, P ¼ 0.013) (Table 1). Under theTPM, the Wilcoxon test revealed significant patterns ofexcess heterozygosity (P(E) ¼ 0.002) (Table 2). Two pairsof genetically identical individuals were identified froma total of 24 collected individuals (Sb 09/13 and Sb 19/21) (Fig. 2). Eight microsatellite loci used in the study ofS. brachyodon population were sufficiently polymorphicto allow individual identification from among 800 × 106

individuals based on the estimate of the unbiased PI(PIunbiased ¼ 1.25 × 1029) or among 1.286 individuals by



by guest on Novem


Dow

nloaded from


assuming that a population was composed only of sibs(PIsibs ¼ 7.77 × 1024). The probabilities of obtaining thesame multilocus genotypes through distinct sexualrecombination events were low in both cases (Psex ¼

5.83 × 1027 and 4.97 × 1028, respectively), and evenwhen using a more conservative estimate [Psex( f )] thattakes into account possible departures from HWE, thetest yielded extremely low values [Psex( f ) ¼ 1.04 × 1026

and 1.25 × 1027, respectively].In comparison with the S. officinalis population from

the same location, no significant differences betweenobserved and expected heterozygosity were found.

In the S. fruticosa, the observed heterozygosity waslower than expected (0.307 vs. 0.351, respectively),which resulted in a significantly positive FIS value (FIS ¼

0.125, P ¼ 0.044). Under the TPM, the Wilcoxon testrevealed a pattern of heterozygosity deficiency (P(D) ¼0.027) (Table 2). When comparing the overall populationparameters of S. fruticosa and sympatric S. officinalis, theS. fruticosa presented significantly lower values of thenumber of alleles, observed and expected heterozygosity(P ¼ 0.012, 0.005 and 0.016, respectively) (Table 1). Theclose genetic relationship between individuals was con-firmed (Fig. 2). In addition to S. officinalis and S. fruticosa,a third group of intermediate individuals was detected,suggesting hybridization between these species (Figs 3and 4). The results from the Bayesian analysis implemen-ted in STRUCTURE confirmed the existence of two separateclusters (i.e. parental species) and a group of three indivi-duals characterized by admixed proportions (Fig. 4). Thehighest DK was obtained at K ¼ 2 (1862.93), while thesecond highest DK was at K ¼ 4 (9.62). The presence oftwo well-differentiated populations belonging to parentaltaxa was strongly supported by the above results, as wasthe existence of individuals of hybrid origin.

The hybrids assignment by NEWHYBRIDS was in con-cordance with the results from the STRUCTURE (Fig. 4).All individuals characterized by admixed proportionswere confirmed as F1 hybrids with very similar posteriorprobabilities ranging from 0.762 to 0.773. Interestingly,

posterior probabilities suggesting the possible origin ofthese individuals through backcrossing were relativelyhigh and ranged from 0.186 to 0.212.

Figure 2. An unrooted Fitch–Margoliash tree based on the propor-tion-of-shared-allele distances among samples of S. brachyodon,S. officinalis and S. fruticosa (from top to bottom, respectively). Boot-strap support values higher than 50 % for 1000 replicates are indi-cated on the branches. Salvia officinalis samples marked as ‘So1-’and ‘So2-’ are from Peljesac Peninsula and the Island of Vis localities,respectively.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Table 2. Probabilities of heterozygote excess [P(E)] and deficiency[P(D)] according to a Wilcoxon test under the TPM assuming 30 %of IAM and 70 % of SMM.

Population Locality TPM TPM

P(D) P(E)

S. brachyodon Peljesac 1.000 0.002

S. officinalis Peljesac 0.680 0.371

S. officinalis Vis 0.320 0.727

S. fruticosa Vis 0.027 0.980



by guest on Novem


Dow

nloaded from


Discussion

S. officinalis

In the S. officinalis from Peljesac, most loci were in HWE,with the exception of locus SoUZ011, which was charac-terized by a significantly positive FIS value (Table 1). Theseresults are consistent with the expectation that thispopulation, located in the central part of the eastern Adri-atic mainland coastal region where common sage growsabundantly, is in mutation-drift equilibrium and is notexperiencing any population genetic disturbances. Theresults for the population from the Island of Vis werevery similar; deviations from HWE were insignificant,and there were no statistically significant population gen-etic differences between these two populations (Table 1).Being located on an offshore island did not cause a sub-stantial disturbance in the genetic structure of this popu-lation. Plausible explanations for this situation may lie inthe fact that some 20 000 years ago, during the Last Gla-cial Maximum, when sea level was �120 m lower thantoday (Rohling et al. 1998; Clark et al. 2009), the Islandof Vis was part of the mainland. After the sea level rose,the remaining population was simply large enough to

retain most of the genetic variability that has been pre-served until the present day. The low level of differenti-ation between the two S. officinalis populations (Fig. 2)is not surprising because the Peljesac Peninsula is thenearest mainland to the Island of Vis.

Being a typical heliophyte and one of the most abun-dant plant species in open habitats along the easternAdriatic coast, S. officinalis is not threatened currently.The selected populations located in the geographicalcentre of the species distribution area are characterizedby high levels of genetic variability and are not experien-cing any population genetic disruptive circumstances.Because the results confirm hypothesis (1), S. officinalispopulations can be treated as reference populations incontrast to populations of the other two species thatprobably experienced some of these disruptive factors.

S. brachyodon

In comparison to the S. officinalis from the same location,S. brachyodon was characterized by no significant differ-ences between the observed and expected levels of het-erozygosity. There are several possible explanations for

Figure 3. A neighborNet diagram based on the proportion-of-shared-alleles distance matrix among individuals belonging to S. officinalis (in blue),S. fruticosa (in orange) and their hybrid, S. × auriculata (in yellow). The bootstrap support value was derived from a Neighbor Joining analysis.



by guest on Novem


Dow

nloaded from


such high heterozygosity levels and heterozygosityexcess (negative FIS value). First, during the samplingexpedition, a large number of bumblebees (Bombus sp.)were observed to visit S. brachyodon flowers, supportingthe expectation that this species, like most sage specieswith large flowers (Haque and Ghoshal 1981), is predom-inantly an outbreeding species. Second, although thelarge and branched inflorescences that usually growover 1 m in height produce a significant number of flow-ers, they tend to bloom sequentially (basitonic type),resulting in very few flowers on the same plant beingavailable for pollination simultaneously (I. Radosavljevic,Z. Satovic and Z. Liber, pers. obs.), thus favouring pollin-ation between different individuals. These two traitsmay help to maintain genetic variation by preventingself-pollination; that is, by promoting outcrossing, theprobability of creating homozygous offspring is reduced.Third, S. brachyodon tends not to grow individually butrather in dense groups of individuals, which indicates clo-nal reproduction by stolons (Fig. 1). Within our sample setof 24 individuals, two sample pairs were shown to be gen-etically identical (Fig. 2), thus supporting our observation.Additional conformation that clonality, rather than highinbreeding levels, is responsible for the presence of indivi-duals sharing the same multilocus genotype, wasobtained by GenClone software. With clonality being

confirmed, the negative FIS value, e.g. heterozygosityexcess, could easily be explained (Delmotte et al. 2002;Balloux et al. 2003; Alberto et al. 2005; Ruggiero et al.2005). As stated several times (Judson and Normark1996; Welch and Meselson 2000), asexual reproductioncan be responsible for maintaining high heterozygositylevels or even for increasing heterozygosity by the accu-mulation of mutations over generations. New mutationsthat occur are permanently fixed and cannot be lostthrough genetic drift because of the presence of non-sexual reproduction.

The BOTTLENECK test provided evidence that thisS. brachyodon population has recently experienced severereductions in the effective population size, i.e. a bottle-neck. There are two possible explanations for a popula-tion bottleneck that should be taken into account:wildfire and ecological succession. Twelve years beforesampling, in August of 1998, a devastating fire overranthe entire area where this population is located (Fig. 1).However, as documented on numerous occasions,through the re-establishment of open, non-shaded habi-tats, periodic fires in Mediterranean-type ecosystems playa vital role in their evolutionary history (Naveh 1975;Keeley 1977; Ne’eman and Dafni 1999), rather than hav-ing a destructive impact on local biodiversity. Neverthe-less, human-induced fires are prevalent in contrast to

Figure 4. Proportions of membership of each individual in each of the two clusters as estimated by the program STRUCTURE. Each individualplant is represented by a single vertical line divided into colours. Each colour represents one cluster, and the length of the coloured segmentshows the individual’s estimated proportion of membership in that cluster. Assignment of individuals into classes (parental S. officinalis andS. fruticosa and F1) based on maximum posterior probabilities that each individual belongs to a particular class as estimated by the programNEWHYBRIDS.



by guest on Novem


Dow

nloaded from


naturally occurring fires, as only 1–5 % of all the wildfiresin the Mediterranean are of natural origin (Alexandrianet al. 1998). Thus, it is very likely that this specific firewas also human-induced. However, the presence of thisisolated population suggests that S. brachyodon, being anative Mediterranean species, is well adapted to such anextreme event and it is unlikely that the fire caused anysignificant contraction in population size. Although thespecies cannot be characterized as a hemicryptophytebut rather as a chamaephyte, due to its clonal reproduc-tion and the presence of underground stolons, it is alsolikely that resprouting is a possible mechanism of post-fire regeneration (Naveh 1975; Keeley 1977; Lloret et al.2003). It is also possible that S. brachyodon is, such asS. fruticosa, an obligate seeder (Ne’eman and Dafni1999), and may be dependent on a soil-stored seedbank. For such species, in the absence of human activitiesthat support open-type habitats (e.g. certain styles ofagriculture), occasional fires are needed to ‘clear’ the ter-rain of vegetation, consequently allowing seedlings toestablish (Ne’eman and Dafni 1999; Santana et al.2012). Once pastures, today the karstic plains in the Pelje-sac Peninsula are mostly abandoned; consequently, grad-ual ecological succession towards the indigenous blackpine (Pinus nigra Arnold) forest (Flora Croatica Database;http://hirc.botanic.hr/fcd) is common and inevitablethroughout the region. Being a typical heliophyte,S. brachyodon requires clear, open space to thrive and theabandonment of the traditional pasture managementmay result in a loss of a suitable habitats. The ecologicalsuccession at this location prior to the 1998 fire is themost likely reason for this population bottleneck and recentwildfire likely saved this population from deterioration andpossibly even extinction. To re-establish and maintain thedesired open habitat and the associated species commu-nity that emerged from the wildfire, grazing cattle couldbe reintroduced in such areas (Verdu et al. 2000).

Although narrow endemic species in general areconsidered to be characterized by low levels of geneticdiversity (e.g. Gitzendanner and Soltis 2000), on someoccasions it has proven to be the opposite (e.g. Dolanet al. 1999; Martinez-Palacios et al. 1999; Hannan andOrick 2000). Therefore, the high levels of genetic variationfound in S. brachyodon that failed to support the pre-dicted hypothesis (2) are not unprecedented. Clonalreproduction can serve as a valuable mechanism to pre-serve genetic diversity and in spite of the detected popu-lation bottleneck, high levels of genetic variability canindeed be found in this species.

S. fruticosa

Two parameters that are interesting for this isolatedpopulation are the positive and highly significant FIS

value, which suggests high inbreeding levels, and the sig-nificant heterozygosity deficiency as revealed by BOTTLE-NECK. The results additionally confirm the hypothesis ofthe non-native origin of S. fruticosa in this location. As isoften the case with anthropogenic introductions, thispopulation was most likely founded from a very limitednumber of individuals who were completely isolatedfrom the source population and from any other popula-tion of their kind. The resulting genetic drift, bottleneckand inevitable breeding between close relatives subse-quently led to increased homozygosity, which had asignificantly higher value when compared with the sym-patric S. officinalis population (Table 1). Furthermore, ifour hypothesis was correct and assuming that this intro-duction was intentional due to the settlers’ (i.e. AncientGreeks) knowledge of the aromatic and health-beneficialproperties of this species (Rivera et al. 1994), it is reason-able to assume that this population was anthropogeni-cally restrained from expansion and was composed of avery limited number of individuals for a long periodof time. During this prolonged bottleneck period,S. fruticosa was grown only in gardens and vineyards sur-rounding local settlements. The expansion of S. fruticosatook place just recently, in approximately the last hun-dred years, when much of the local population emigrated,resulting in the abandonment of traditional agriculture.This hypothesis is indeed consistent with the current dis-tribution of S. fruticosa on the Island of Vis because it isfound exclusively in open, non-shaded habitats ofneglected and abandoned vineyards and olive groves sur-rounding local settlements. The fact that S. fruticosa is anative East-Mediterranean species has eased its adapta-tion to a new environment that did not differ significantlyfrom that in its native area of distribution, allowingS. fruticosa to expand quickly and form a current popula-tion of a few thousand individuals. In addition, the closegenetic relationship within the population that can beobserved today (Fig. 2) supports the hypothesis that thispopulation was started relatively recently by a very limitednumber of individuals. During this relatively short period,there was not enough time, i.e. not enough generations,for newly emerging mutations to spread through thepopulation and to accumulate in higher frequencies.Thus, we are now observing a significant heterozygositydeficiency (HE , HEQ). It is likely that over time and withina population of this size, the frequency of new alleles, andconsequently heterozygosity, will increase.

However, considering all the results obtained in thisstudy, it is becoming clearer that favourable ecologicalparameters are not the only factors responsible for sucha vigorous and successful expansion of S. fruticosa. Asalready mentioned, S. officinalis and S. fruticosa occursympatrically in this location with overlapping flowering



by guest on Novem


Dow

nloaded from





periods. It is also important to note that a putative hybridbetween these species has been recorded (Salvia × auri-culata Mill.) (Putievsky et al. 1990), but to the best of ourknowledge, only as a result of artificial crossing (Putievskyet al. 1990; Dudai et al. 1999; Reales et al. 2004). It is alsonoteworthy that three sampled specimens of S. fruticosawere characterized by non-typical morphological traitsfor this species, especially the calyx shape (non-radial,slightly zygomorphic) and the calyx trichome lengthand density (which were more pronounced than usual).However, because this species is characterized by excep-tionally high levels of morphological variability (Hedge1972; Karousou and Kokkini 1997), these samples wereconsidered to originate from S. fruticosa. As revealed bythe computer programs SplitsTree (Fig. 3), STRUCTUREand NEWHYBRIDS (Fig. 4) all three doubtful individuals(Sf22, Sf23 and Sf24) were of hybrid origin. Although theresults undoubtedly support the origin of these indivi-duals through hybridization, their asymmetric positioningin the neighbour-net diagram (Fig. 3) and somewhatinconclusive results from NEWHYBRIDS (Fig. 4) also sug-gest their possible origin through backcrossing with oneof the parental species, i.e. S. fruticosa. However, the sam-ple set is not large enough for more detailed and compre-hensive study of the hybridization event, so it isimpossible to address some very important questions,such as (i) are the hybrids fertile, (ii) is backcrossing toeither parental species indeed present and (iii) if back-crossing is present, is it symmetrical or unilateral etc. Ashybridization between species is generally considered tobe rare on an individual basis, a very limited number ofindividuals participate in hybridization events. Conse-quently, hybrids are rare in the population, but even afew hybrids can provide a bridge that allows a trickle ofalleles to pass between species (Mallet 2005). Thus, ifwe leave the possibility of the backcrossing withS. fruticosa open, the appearance of certain low-frequency alleles in S. fruticosa is not surprising. As newalleles originating from S. officinalis reach S. fruticosathrough backcrossing, the alleles can be found in only asmall percentage of individuals. Consequently, the fre-quency of these genes in the new gene pool is very low.They therefore contribute considerably to increased alle-lic diversity but not to heterozygosity, which is conse-quently detected by BOTTLENECK as heterozygositydeficiency (Table 2). However, as S. officinalis andS. fruticosa are undoubtedly closely related species it ispossible to assume that at least some of these rare com-mon alleles are not necessarily the result of inter-speciesintrogression and horizontal gene transfer, but insteadshare the same origin through a common ancestor. It isreasonable to assume that the high adaptability as wellas the genetic and range expansion that can be observed

in the S. fruticosa population is supported by introgressionand hybridization with a successful and well-adaptedmember of the indigenous flora, S. officinalis.

In addition, naturally occurring hybridization betweenthese species may be of significant agricultural interest.As mentioned above, artificial crossing between thesespecies has been already performed, and comparativeresearch should address the yield and content of essentialoil from hybrid individuals of artificial and natural origin.

The results support the hypothesis (3) that theS. fruticosa population is characterized by very low levelsof genetic variability. As a bottleneck event can be detectedin a limited time frame after it happens (Peery et al. 2012),we were unable to recognize it. However, high levels ofinbreeding accompanied by very low levels of genetic vari-ability suggested the occurrence of a bottleneck event inthe past. The inter-population gene flow, expected to beabsent in remote and isolated populations, was possiblypresent, but on an inter-species level. Hybridization withindigenous S. officinalis was confirmed, but its nature anddirection are still uncertain.

ConclusionsOur initial prediction regarding the levels of genetic diver-sity of three sympatrically growing Salvia species wereonly partially confirmed. In accordance with expecta-tions, S. officinalis populations, as true core populations,were characterized by high levels of variability, whilethe disjunct and isolated population of S. fruticosa con-tained significantly less genetic variation and struggledwith high levels of inbreeding. The natural hybridizationbetween S. officinalis and S. fruticosa was confirmed forthe first time. To assess the nature and direction of hybrid-ization, more extensive research including a significantlylarger number of samples and additional methods shouldbe conducted. Although morphometrical analyses aloneare considered to be of limited value in hybrid identificationand characterization (Rieseberg et al. 1993), if combinedwith molecular methods, e.g. SSRs, more informative andinsightful results could be obtained.

In contrast to expectation, the population of theendangered and narrow endemic S. brachyodon, evenafter surviving a population bottleneck, was still charac-terized by surprisingly high levels of genetic diversity.Detected clonal reproduction via underground stolonscould serve as a valuable mechanism and adaptation inpreservation and even accumulation of emerging muta-tions, although a larger sample set and appropriateanalysis methods are needed for more comprehensiveresults and conclusions. Detailed spatial-genetic analysiswould give an inside perspective on this crucial life trait(i.e. clonality) as numerous questions dealing with this



by guest on Novem


Dow

nloaded from


species’ historical and contemporary demography couldbe answered.

Some of the detected fluctuations in genetic structuresof the studied Salvia populations could also be a result ofcertain human activities or their absence. However, beforereaching any conclusions, caution is needed, as somedetected genetic traits could easily be a result of specificdifferences in species life history traits, rather than a con-sequence of human activities. To more reliably address thisand similar topics, a different approach is needed.

This research demonstrates the importance and advan-tages of a multi-species approach in population geneticsstudies, as some relevant findings and conclusions abouteach species can be reached only if a comparison amongdifferent, yet similar and closely related species is per-formed. The selection of the ‘control’ group should becarefully considered, as it will set the frame of the entireresearch. Presumably, widespread species with no signifi-cant population genetic disturbances, highly specific lifetraits or large-scale demographic fluctuations, should beconsidered as the right choice for such a purpose.

Sources of FundingThis study was supported by the Croatian Science Foun-dation within the framework of the Project No. 09.01/246 (Epigenetic vs genetic diversity in natural plant popu-lations: A case study of Croatian endemic Salvia species).

Contributions by the AuthorsI.R. performed the experiment, wrote and edited themanuscript and created all the figures. Z.S. conducted stat-istical analyses. Z.L. conceived the idea for the researchand extensively edited the manuscript. All authors wereequally involved in sample collection.

Conflict of Interest StatementNone declared.

AcknowledgementsThe authors thank two anonymous reviewers and anAssociate Editor for their useful comments that substan-tially improved this manuscript. Our sincere gratitude isalso expressed to our colleague, Bostjan Surina, for hiscomments on the previous versions of the manuscript.

Literature CitedAbadzic S, Silic C. 1982. Chorology, ecology and phytosociological

affiliation of Salvia brachyodon Vandas in flora of Yugoslavia (inMontenegrin). Glasnik Republickog Zavoda za Zastitu Prirode i Prir-odnjackog Muzeja u Titogradu 15:125–131.

Alberto F, Gouveia L, Arnaud-Haond S, Perez-Llorens JL, Duarte CM,Serrao EA. 2005. Within-population spatial genetic structure,neighbourhood size and clonal subrange in the seagrass Cymo-docea nodosa. Molecular Ecology 14:2669–2681.

Alexandrian D, Esnault F, Calabri G. 1998. Forest fires in the Mediter-ranean area. Food and Agriculture Organization of the UnitedNations, The Forestry Department.

Anderson EC, Thompson EA. 2002. A model-based method for iden-tifying species hybrids using multilocus genetic data. Genetics160:1217–1229.

Arnaud-Haond S, Belkhir K. 2007. GENCLONE 1.0: a new program toanalyze genetics data on clonal organism. Molecular EcologyNotes 7:15–17.

Arnaud-Haond S, Duarte CM, Alberto F, Serrao EA. 2007. Standardiz-ing methods to address clonality in population studies. MolecularEcology 16:5115–5139.

Arnold ML. 1997. Natural hybridization and evolution. Oxford: OxfordUniversity Press.

Balloux F, Lehmann L, de Meeus T. 2003. The population genetics ofclonal and partially clonal diploids. Genetics 164:1635–1644.

Barbalic Lj. 1956. Contribution to the knowledge of Salvia brachyodonVand (in Croatian). Periodicum Biologorum 9:5–10.

Blondel J. 2006. The ‘design’ of Mediterranean landscapes: a millen-nial story of humans and ecological systems during the historicperiod. Human Ecology 34:713–729.

Bowcock AM, Ruiz-Linares A, Tomfohrde J, Minch E, Kidd JR, Cavalli-Sforza LL. 1994. High resolution of human evolutionary trees withpolymorphic microsatellites. Nature 368:455–457.

Clark PU, Dyke AS, Shakun JD, Carlson AE, Clark J, Wohlfarth B,Mitrovica JX, Hostetler SW, McCabe AM. 2009. The last glacialmaximum. Science 325:710–714.

Cornuet JM, Luikart G. 1996. Description and power analysis of twotests for detecting recent population bottlenecks from allele fre-quency data. Genetics 144:2001–2014.

Cutini M, Cancellieri L, Ceschin S, Lucchese F, Caneva G. 2007. Analisicenologica e sintassonomia delle garighe a Salvia officinalisL. lucane nel quadro dei salvieti peninsulari (Basilicata, Appen-nino meridionale). Webbia 62:225–244.

Delmotte F, Leterme N, Gauthier J-P, Rispe C, Simon J-C. 2002. Gen-etic architecture of sexual and asexual populations of the aphidRhopalosiphum padi based on allozyme and microsatellite mar-kers. Molecular Ecology 11:711–723.

Di Pietro R. 2011. New dry grassland associations from the Ausoni-Aurunci mountains (central Italy)—syntaxonomical updatingand discussion on the higher rank syntaxa. Hacquetia 10:183–231.

Di Rienzo A, Peterson AC, Garza JC, Valdes AM, Slatkin M, Freimer NB.1994. Mutational processes of simple-sequence repeat loci inhuman populations. Proceedings of the National Academy ofSciences of the USA 91:3166–3170.

Dolan RW, Yahr R, Menges ES, Halfhill MD. 1999. Conservationimplications of genetic variation in three rare species endemicto Florida rosemary scrub. American Journal of Botany 86:1556–1562.

Dudai N, Lewinsohn E, Larkov O, Katzir I, Ravid U, Chimovitsh D,Sa’ad D, Putievsky E. 1999. Dynamics of yield components andessential oil production in a commercial hybrid sage (Salviaofficinalis × Salvia fruticosa cv. Newe Ya’ar No. 4). Journal of Agri-cultural and Food Chemistry 47:4341–4345.



by guest on Novem


Dow

nloaded from


Earl DA, vanHoldt BM. 2012. STRUCTURE HARVESTER: a website andprogram for visualizing STRUCTURE output and implementingthe Evanno method. Conservation Genetics Resources 4:359–361.

Eckert CG, Samis KE, Lougheed SC. 2008. Genetic variation acrossspecies’ geographical ranges: the central–marginal hypothesisand beyond. Molecular Ecology 17:1170–1188.

Evanno G, Regnaut S, Goudet J. 2005. Detecting the number of clus-ters of individuals using the software STRUCTURE: a simulationstudy. Molecular Ecology 14:2611–2620.

Evett I, Weir B. 1998. Interpreting DNA evidence: statistical geneticsfor forensic scientists. Sunderland: Sinauer Associates.

Flora Croatica Database. http://hirc.botanic.hr/fcd (21 May 2013).

Fonderflick J, Lepart J, Caplat P, Debussche M, Marty P. 2010. Man-aging agricultural change for biodiversity conservation in a Medi-terranean upland. Biological Conservation 143:737–746.

Gaston KJ. 2003. The structure and dynamics of geographic ranges.Oxford: Oxford University Press.

Gitzendanner MA, Soltis PS. 2000. Patterns of genetic variation in rareand widespread plant congeners. American Journal of Botany 87:783–792.

Hannan GL, Orick MW. 2000. Isozyme diversity in Iris cristata and thethreatened glacial endemic I. lacustris (Iridaceae). AmericanJournal of Botany 87:293–301.

Haque MS, Ghoshal KK. 1981. Floral biology and breeding system inthe genus Salvia L. Proceedings of the National Academy ofSciences, India 47:716–724.

Hedge IC. 1972. Salvia. In: Tutin TG, Heywood VH, Burges NA,Moore DM, Valentine DH, Walters SM, Webb DA, eds. Flora Euro-paea, Vol. 3. Cambridge: Cambridge University Press, 188–192.

Hedge IC. 1982. Salvia L. In: Davis PH, ed. Flora of Turkey and the EastAegean Islands, Vol. 7. Edinburgh: Edinburgh University Press,400–461.

Hewitt GM. 1999. Post-glacial re-colonization of European biota.Biological Journal of the Linnean Society 68:87–112.

Hewitt GM. 2001. Speciation, hybrid zones and phylogeography—orseeing genes in space and time. Molecular Ecology 10:537–549.

Huson DH, Bryant D. 2006. Application of phylogenetic networks inevolutionary studies. Molecular Biology and Evolution 23:254–267.

Judson OP, Normark BB. 1996. Ancient asexual scandals. Trends inEcology and Evolution 11:41–46.

Karousou R, Kokkini S. 1997. Distribution and clinal variation of Salviafruticosa Mill. (Labiatae) on the island of Crete (Greece). Willdeno-wia 27:113–120.

Karousou R, Hanlidou E, Kokkini S. 2000. The Sage plants of Greece:distribution and infraspecific variation. In: Kintzios SE, ed. SAGEThe genus Salvia. Amsteldijk: Harwood Academic Publishersimprint, part of the Gordon and Breach Publishing Group, 27–46.

Keeley JE. 1977. Seed production, seed populations in soil, and seed-ling production after fire for two congeneric pairs of sproutingand nonsprouting chaparal shrubs. Ecology 58:820–829.

Lesica P, Allendorf FW. 1995. When are peripheral populations valu-able for conservation? Conservation Biology 9:753–760.

Lloret F, Pausas JG, Vila M. 2003. Responses of Mediterranean PlantSpecies to different fire frequencies in Garraf Natural Park (Cata-lonia, Spain): field observations and modelling predictions. PlantEcology 167:223–235.

Lucchese F, Persia G, Pignatti S. 1995. I prati a Bromus erectus Hud-son dell’Appennino Laziale. Fitosociologia 30:145–180.

Luikart G, Allendorf FW, Cornuet J-M, Sherwin WB. 1998. Distortion ofallele frequency distributions provides a test for recent popula-tion bottlenecks. Journal of Heredity 89:238–247.

Mallet J. 2005. Hybridization as an invasion of the genome. Trends inEcology and Evolution 20:229–237.

Martinez-Palacios A, Eguiarte LE, Furnier GR. 1999. Genetic diversityof the endangered endemic Agave victoriae-reginae (Agavaceae)in the Chihuahuan desert. American Journal of Botany 86:1093–1098.

Meeus S, Honnay O, Jacquemyn H. 2012. Strong differences in gen-etic structure across disjunct, edge, and core populations of thedistylous forest herb Pulmonaria officinalis (Boraginaceae).American Journal of Botany 99:1809–1818.

Michaux JR, Magnanou E, Paradis E, Nieberding C, Libois R. 2003.Mitochondrial phylogeography of the woodmouse (Apodemussylvaticus) in the western Palearctic region. Molecular Ecology12:685–697.

Miller NF. 1992. The origins of plant cultivation in the Near East. In:Cowan CW, Watson PJ, eds. The origins of agriculture: an inter-national perspective. Washington: Smithsonian Institution Press,39–58.

Minch E, Ruiz-Linares A, Goldstein D, Feldman M, Cavalli-Sforza LL.1996. Microsat: a computer program for calculating various statis-tics on microsatellite allele data. Version 1.5. Stanford: StanfordUniversity Medical Center.

Molecular Ecology Resources Primer Development Consortium, An J,Bechet A, Berggren A, Brown SK, Bruford MW, Cai Q,Cassel-Lundhagen A, Cezilly F, Chen S-L, Cheng W, Choi S-K,Ding XY, Fan Y, Feldheim KA, Feng ZY, Friesen VL, Gaillard M,Galaraza JA, Gallo L, Ganeshaiah KN, Geraci J, Gibbons JG.,Grant WS., Grauvogel Z, Gustafsson S, Guyon JR., Han L,Heath DD, Hemmila S, Hogan JD, Hou BW, Jakse J, Javornik B,Kanuch P, Kim K-K, Kim K-S, Kim S-G, Kim S-I, Kim W-J, Kim Y-K,Klich MA, Kreiser BR, Kwan Y-S, Lam AW, Lasater K, Lascoux M,Lee H, Lee Y-S, Li DL, Li S-J, Li WY, Liao X, Liber Z, Lin L, Liu S,Luo X-H, Ma YH, Ma Y, Marchelli P, Min M-S, Moccia MD,Mohana KP, Moore M, Morris-Pocock JA, Park H-C, Pfunder M,Ivan R, Ravikanth G, Roderick GK., Rokas A, Sacks BN, Saski CA,Satovic Z, Schoville SD, Sebastiani F, Sha Z-X, Shin E-H, Soliani C,Sreejayan N, Sun Z, Tao Y, Taylor SA, Templin WD, Shaanker RU,Vasudeva R, Vendramin GG, Walter RP, Wang G-Z, Wang K-J,Wang YQ, Wattier RA, Wei F, Widmer A, Woltmann S, Won Y-J,Wu J, Xie ML, Xu G, Xu X-J, Ye H-H, Zhan X, Zhang F, Zhong J.2010. Permanent genetic resources added to molecular ecologyresources database 1 October 2009–30 November 2009. Molecu-lar Ecology Resources 10:404–408.

Naveh Z. 1975. The evolutionary significance of fire in the Mediterra-nean region. Vegetatio 29:199–208.

Naveh Z, Dan J. 1973. The human degradation of Mediterraneanlandscapes in Israel. In: di Castri F, Mooney HA, eds. Mediterra-nean type ecosystems: origin and structure. Berlin: Springer,372–390.

Ne’eman G, Dafni A. 1999. Fire, bees, and seed production in a Medi-terranean key species Salvia fruticosa Miller (Lamiaceae). IsraelJournal of Plant Sciences 47:157–163.

Parks JC, Werth CR. 1993. A study of spatial features of clones in apopulation of Bracken fern, Pteridium aquilinum (Dennstaedtia-ceae). American Journal of Botany 80:537–544.

Peery MZ, Kirby R, Reid BN, Stoelting R, Doucet-Beer E, Robinson S,Vasquez-Carrillo C, Pauli JN, Palsbøll PJ. 2012. Reliability of



by guest on Novem


Dow

nloaded from





genetic bottleneck tests for detecting recent population declines.Molecular Ecology 21:3403–3418.

Petrovic D, Stesevic D, Vuksanovic S. 2008. Materials for the Red bookof Montenegro. Natura Montenegrina 7:605–631.

Pignatti S. 1982. Flora d’ Italia, Vol. 2. Bologna: Edagricole.

Piry S, Luikart G, Cornuet JM. 1999. BOTTLENECK: a computer pro-gram for detecting recent reductions in the effective populationsize using allele frequency data. Journal of Heredity 90:502–503.

Pritchard JK, Stephens M, Donnelly P. 2000. Inference of populationstructure using multilocus genotype data. Genetics 155:945–959.

Putievsky E, Ravid U, Diwan-Rinzler N, Zohary D. 1990. Genetic affin-ities and essential oil composition of Salvia officinalis L.,S. fruticosa Mill., S. tomentosa Mill. and their hybrids. Flavourand Fragrance Journal 5:121–123.

Radosavljevic I, Jakse J, Javornik B, Satovic Z, Liber Z. 2011. Newmicrosatellite markers for Salvia officinalis (Lamiaceae) andcross-amplification in closely related species. American Journalof Botany 98:e316–e318.

Radosavljevic I, Satovic Z, Jakse J, Javornik B, Greguras D,Jug-Dujakovic M, Liber Z. 2012. Development of new microsatel-lite markers for Salvia officinalis L. and its potential use inconservation-genetic studies of narrow endemic Salvia brachyo-don Vandas. International Journal of Molecular Sciences 13:12082–12093.

Raymond M, Rousset F. 1995. GenePop (version 1.2): population gen-etics software for exact tests and ecumenicism. Journal of Hered-ity 86:248–249.

Reales A, Riviera D, Palazon JA, Obon C. 2004. Numerical taxonomystudy of Salvia sect. Salvia (Labiatae). Botanical Journal of the Lin-nean Society 145:353–371.

Rieseberg LH, Ellstrand NC, Arnold M. 1993. What can molecular andmorphological markers tell us about plant hybridization? CriticalReviews in Plant Sciences 12:213–241.

Rieseberg LH, Raymond O, Rosenthal DM, Lai Z, Livingstone K,Nakazato T, Durphy JL, Schwarzbach AE, Donovan LA, Lexer C.2003. Major ecological transitions in wild sunflowers facilitatedby hybridization. Science 301:1211–1216.

Rivera D, Obon C, Cano F. 1994. The botany, history and traditionaluses of three-lobed sage (Salvia fruticosa Miller) (Labiatae). Eco-nomic Botany 48:190–195.

Rohling EJ, Fenton M, Jorissen FJ, Bertrand P, Ganssen G, Caulet JP.1998. Magnitudes of sea-level lowstands of the past 500,000years. Nature 394:162–165.

Ruggiero MV, Reusch TBH, Procaccini G. 2005. Local genetic structurein a clonal dioecious angiosperm. Molecular Ecology 14:957–967.

Sagarin RD, Gaines SD, Gaylord B. 2006. Moving beyond assumptionsto understand abundance distributions across the ranges of spe-cies. Trends in Ecology and Evolution 21:524–530.

Samis KE, Eckert CG. 2007. Testing the abundant center model usingrange-wide demographic surveys of two coastal dune plants.Ecology 88:1747–1758.

Santana VM, Baeza MJ, Maestre FT. 2012. Seedling establishmentalong post-fire succession in Mediterranean shrublands domi-nated by obligate seeders. Acta Oecologica 39:51–60.

Savikin-Fodulovic KP, Tasic SR, Menkovic NR. 2002. The essential oil ofSalvia brachyodon Vandas. Lamiaceae. Journal of Essential OilResearch 14:342–343.

Seehausen O. 2004. Hybridization and adaptive radiation. Trends inEcology and Evolution 19:198–207.

Silic C. 1984. Endemic plants (in Bosnian). Sarajevo: Svjetlost.

Stebbins GL. 1959. The role of hybridization in evolution. Proceedingsof the American Philosophical Society 103:231–251.

Surina B, Schneeweiss GM, Glasnovic P, Schonswetter P. 2014. Testingthe efficiency of nested barriers to dispersal in the Mediterraneanhigh mountain plant Edraianthus graminifolius (Campanula-ceae). Molecular Ecology 23:2861–2875.

Taberlet P, Luikart G. 1999. Non-invasive genetic sampling and indi-vidual identification. Biological Journal of the Linnean Society 68:41–55.

Taberlet P, Fumagalli L, Wust-Saucy A-G, Cosson J-F. 1998. Compara-tive phylogeography and postglacial colonization routes inEurope. Molecular Ecology 7:453–464.

Tzakou O, Couladis M, Slavkovska V, Mimica-Dukic N, Jancic R. 2003.The essential oil composition of Salvia brachyodon Vandas.Flavour and Fragrance Journal 18:2–4.

Valiere N. 2002. GIMLET: a computer program for analysing geneticindividual identification data. Molecular Ecology Notes 2:377–379.

van Oosterhout C, Hutchinson WF, Wills DPM, Shipley P. 2004. Micro-Checker: software for identifying and correcting genotypingerrors in microsatellite data. Molecular Ecology Notes 4:535–538.

Vasic O. 1997. A survey of the Mediterranean species of Lamiaceaefamily in the flora of Serbia. Lagascalia 19:263–270.

Verdu JR, Crespo MB, Galante E. 2000. Conservation strategy of anature reserve in Mediterranean ecosystems: the effects of pro-tection from grazing on biodiversity. Biodiversity and Conserva-tion 9:1707–1721.

Waits LP, Luikart G, Taberlet P. 2001. Estimating the probability ofidentity among genotypes in natural populations: cautions andguidelines. Molecular Ecology 10:249–256.

Walker JB, Sytsma KJ, Treutlein J, Wink M. 2004. Salvia (Lamiaceae) isnot monophyletic: implications for the systematics, radiation,and ecological specializations of Salvia and tribe Mentheae.American Journal of Botany 91:1115–1125.

Welch DM, Meselson M. 2000. Evidence for the evolution of bdelloidrotifers without sexual reproduction or genetic exchange.Science 288:1211–1215.

Zeder MA. 2008. Domestication and early agriculture in the Mediter-ranean Basin: origins, diffusion, and impact. Proceedings of theNational Academy of Sciences of the USA 105:11597–11604.

Zeder MA, Hesse B. 2000. The initial domestication of goats (Caprahircus) in the Zagros Mountains 10,000 years ago. Science 287:2254–2257.



by guest on Novem


Dow

nloaded from


Documents

Causes and consequences of contrasting genetic structure in sympatrically growing …hirc.botanic.hr/EpiSalvia/Docs/Radosavljevic_2015_AOB... · 2015. 11. 15. · tions have been